Depressive symptom severity as a predictor of attendance in the HOME behavioral weight loss trial

Abstract

Objective: We examined whether total depressive symptoms and symptom clusters predicted behavioral weight loss attendance among economically disadvantaged adults in a randomized controlled trial.

Methods: 150 adults with obesity were randomized to 12 months of in-person, video conference, or enhanced usual care weight loss groups. We categorized percent session attendance in the intervention arms into three levels: no attendance, poorer attendance, and better attendance.

Results: Higher baseline Patient Health Questionnaire-8 (PHQ-8) score was associated with a greater odds of being in the poorer versus better attendance group (OR = 1.94, 95% CI: 1.02-3.69, p = .04). A similar relationship between PHQ-8 score and odds of being in the no attendance versus better attendance group was observed but was not statistically significant (OR = 1.63, 95% CI: 0.94-2.81, p = .08). Both cognitive/affective and somatic clusters contributed to the depressive symptoms-attendance relationships.

Conclusion: Greater depressive symptoms at the start of a behavioral weight loss program may predict poorer subsequent session attendance. Screening for and addressing depression may improve intervention uptake. ClinicalTrials.gov Identifier: NCT02057952.

Keywords: Attendance; Clinical trial; Depression; Intervention; Obesity; Weight loss.

Conflict of interest statement

Declaration of Competing Interest The authors have no competing interest to report.

Copyright © 2020 Elsevier Inc. All rights reserved.

Figures

Figure 1.

Results of multinomial logistic regression models examining the associations of total depressive symptom severity (PHQ-8 total z score) and depressive symptom clusters (PHQ-8 cognitive/affective and somatic z scores) with behavioral weight loss session attendance (no attendance, poorer attendance, better attendance). No attendance, poorer attendance, and better attendance groups were defined as percent of total sessions attended equal to 0% (n = 44), 0.1–32.9% (n = 24), and ≥ 33.0% (n = 26), respectively. Models included the 94 selected participants and were adjusted for age, sex, race, waist circumference, and health literacy. PHQ-8 = Patient Health Questionnaire-8. OR = odds ratio. CI = confidence interval. *p < .05